xDage Creations

KMV Credit Portfolio Model & Expected Default Frequency (EDF™) Lab

Simulate firm asset values, Distance to Default ($DD$), empirical Expected Default Frequency ($EDF$), and multi-obligor portfolio CreditVaR using Kealhofer-McQuown-Vasicek structural credit dynamics.

KMV PortfolioManager Moody's Analytics EDF™ Merton Asset Inversion Component Unexpected Loss CreditVaR & CVaR
Portfolio EAD
$170.0M
Total Exposure
Expected Loss
$1.45M
Analytical Mean
Unexpected Loss
$4.85M
Volatility $\sigma_{\text{port}}$
99% CreditVaR
$18.20M
1-Year Horizon
99% CVaR / ES
$22.40M
Tail Shortfall
Economic Capital
$16.75M
$\text{VaR}_{99} - \text{EL}$
Obligor Book & Merton Asset Inversion ($V_A, \sigma_A, DD, EDF$)
Average DD: 2.75 σ
Obligor & Sector Equity ($M) $\sigma_E$ (%) Debt: STD / LTD Inverted $V_A$ & $\sigma_A$ Distance to Default Empirical EDF™ EAD ($M) LGD (%) Component UL Action
Note: Default Point $D = \text{STD} + 0.5 \times \text{LTD}$. Inversion solves Merton system simultaneously for market value of assets $V_A$ and asset volatility $\sigma_A$.
Empirical Moody's KMV EDF™ vs Theoretical Normal $\Phi(-DD)$ Distribution

Notice how the Gaussian standard normal CDF $\Phi(-DD)$ severely understates real-world default probabilities for high-grade credits ($DD > 2.5$) due to empirical fat tails and liquidity freezes.

Simulated Multi-Obligor Portfolio Loss Distribution ($10,000$ Scenarios)

Monte Carlo simulated portfolio credit losses with colored quantile thresholds: Blue (Normal Loss Range), Amber (95% VaR Tail), Red (99% Extreme Tail).

Obligor Risk Attribution: EAD vs Expected Loss (EL) vs Component Unexpected Loss (CUL)

Component Unexpected Loss (CUL) isolates each obligor's marginal risk contribution to portfolio volatility, taking into account asset correlation and default clustering.

Sensitivity: Distance to Default ($DD$) vs Asset Volatility ($\sigma_A$) on EDF (%)
Shows empirical default rate amplification across varying asset volatility environments.
Sensitivity: Asset Correlation ($\rho$) vs EAD Scale on 99% CreditVaR ($M)
Demonstrates the non-linear tail risk expansion as macroeconomic systemic correlation increases.

The Kealhofer-McQuown-Vasicek (KMV) Credit Risk Architecture

In the early 1990s, Stephen Kealhofer, John McQuown, and Oldrich Vasicek founded KMV Corporation (acquired by Moody's in 2002) to bridge the gap between academic structural credit theory (Merton, 1974) and empirical debt market realities. The resulting KMV Model and PortfolioManager™ framework revolutionized commercial banking, fixed income portfolio management, and loan syndication.

1. Structural Asset Inversion

A company's market value of assets $V_A$ and asset volatility $\sigma_A$ are not directly observable. Using Black-Scholes-Merton option pricing theory where equity is modeled as a European call option on firm assets: $$E = V_A \Phi(d_1) - D e^{-rT} \Phi(d_2)$$ $$\sigma_E = \frac{V_A \Phi(d_1)}{E} \sigma_A$$ KMV numerically inverts this non-linear bivariate system to extract true underlying asset values and asset volatilities.

2. The KMV Default Point

Classical theory assumes default occurs when assets drop below total debt. However, KMV empirical research proved that firms do not default immediately when long-term liabilities exceed assets if maturities are distant. Insolvency occurs when liquid resources fail to cover short-term liabilities plus half long-term debt: $$D = \text{STD} + 0.5 \times \text{LTD}$$ This empirical threshold dramatically improved default prediction accuracy.

3. Non-Parametric Empirical EDF™

Distance to Default ($DD$) standardizes how many standard deviations asset values exceed the default point: $$DD = \frac{\ln(V_A / D) + (\mu_A - 0.5\sigma_A^2)T}{\sigma_A \sqrt{T}}$$ Instead of assuming normal distribution tail probabilities $\Phi(-DD)$, KMV maps $DD$ to an empirical database of tens of thousands of corporate defaults, capturing true real-world fat tails.

Institutional Knowledge Check: KMV Framework
Frequently Asked Questions

Developed by Stephen Kealhofer, John McQuown, and Oldrich Vasicek (KMV Corporation, now Moody's Analytics), the KMV model is a market-based structural credit framework. It inverts equity market capitalizations and equity volatilities to determine a firm's unobservable market value of assets and asset volatility. It calculates Distance-to-Default (DD) as the number of standard deviations asset values exceed the default point (Short-Term Debt + 50% Long-Term Debt), and maps this metric to an empirical Expected Default Frequency (EDF) calibrated against large historical default databases.

KMV's extensive empirical research demonstrated that corporations rarely default simply because total liabilities exceed current asset values, provided debt maturities lie far in the future. Instead, insolvency is triggered when cash flows and liquidity are inadequate to service current operating liabilities and maturing debt. KMV found that a threshold equal to short-term liabilities plus half of long-term debt best mirrors actual corporate default boundaries in practice.

The classical Merton (1974) model assumes asset returns follow a standard lognormal random walk, leading to default probability equal to the standard normal CDF Phi(-DD). However, real-world corporate asset returns exhibit heavy negative skewness and leptokurtosis (fat tails). For safe firms with Distance-to-Default greater than 3.5 sigma, the normal distribution predicts default chances near 0.001%, whereas historical default rates are closer to 0.2% to 0.5% due to operational shocks, accounting fraud, and sudden liquidity crises.

KMV PortfolioManager models multi-obligor credit portfolios by decomposing obligor asset returns into systematic global, regional, and industry factors plus idiosyncratic firm shocks. Pairwise asset correlations dictate joint Gaussian return realizations. In each simulation scenario, if an obligor's asset return falls below its standardized default threshold Phi^{-1}(EDF), a default occurs. The resulting portfolio loss distribution yields percentile CreditVaR, Expected Shortfall, and Component Unexpected Loss (CUL).

Component Unexpected Loss (CUL) measures the marginal contribution of an individual loan or exposure to total portfolio unexpected loss: CUL_i = Cov(L_i, L_port) / sigma_port. Because the sum of all individual CUL values exactly equals total portfolio unexpected loss, risk managers use CUL to allocate economic capital, identify concentration risks, and compute Risk-Adjusted Return on Capital (RAROC) for loan pricing.
Explore Quantitative Credit & Portfolio Risk Simulations
CreditRisk+ Actuarial Lab

Simulate portfolio loss distributions, stochastic Poisson default clustering, and CreditVaR using CSFP's actuarial model.

CreditMetrics Portfolio Lab

Simulate multi-obligor credit risk, 8-state rating migrations, asset correlation, and CreditVaR using J.P. Morgan's framework.

CreditPortfolioView Lab

Model macroeconomic credit cycles, stress-tested rating migration matrices, and Wilson logit index multi-factor shocks.

Vasicek Basel IRB Portfolio Lab

Explore Vasicek's Asymptotic Single Risk Factor (ASRF) model, regulatory capital formulas, and 99.9% credit risk quantiles.